DataViz

Practice

Guided Practice

For this section, we are going to use a Rnotebook, which mix both markdwon formatted text and code. We can easily compile the output as a report in .html format to share with others.

We are going to build some cool visualizations using what we talked in the previous section, and this will be the base for the shiny apps in next section.

First of all, we need to load some libraries.

library(tidyverse)
library(gapminder)
library(palmerpenguins)
library(countrycode)
library(ggrepel)
library(plotly)
library(ggthemes)
library(gridExtra)
library(scales)
library(ggridges)

Penguins

penguins %>% 
  filter(!is.na(sex)) %>% 
ggplot(., aes(x= bill_length_mm, y = sex, fill=species)) + 
  geom_density_ridges(alpha=0.5)+
  theme_minimal()+
  scale_fill_viridis_d(na.translate = FALSE)+
  facet_grid(species~.)

penguins %>% 
  filter(!is.na(sex)) %>% 
ggplot(., aes(x = bill_length_mm, y = sex,fill = factor(stat(quantile)))) +
  stat_density_ridges(
    geom = "density_ridges_gradient", calc_ecdf = TRUE,
    quantiles = 4, quantile_lines = TRUE,
    jittered_points = TRUE,
    position = position_points_jitter(width = 0.05, height = 0),
    point_shape = '|', point_size = 3, point_alpha = 1) +
  scale_fill_viridis_d(name = "Quartiles")+
    theme_minimal()+
  facet_grid(species~.)+
  theme(legend.position = 'bottom')

p1 <- penguins %>% 
  filter(!is.na(sex)) %>% 
ggplot(., aes(x = bill_length_mm, y = sex,fill = factor(stat(quantile)))) +
  stat_density_ridges(
    geom = "density_ridges_gradient", calc_ecdf = TRUE,
    quantiles = 4, quantile_lines = TRUE,
    jittered_points = TRUE,
    position = position_points_jitter(width = 0.05, height = 0),
    point_shape = '|', point_size = 3, point_alpha = 1) +
  scale_fill_viridis_d(name = "Quartiles")+
    theme_fivethirtyeight()+
  theme(legend.position = 'none')+
  labs(title = 'Bill lenght') + 
  facet_grid(species~.)

p2 <- penguins %>% 
  filter(!is.na(sex)) %>% 
ggplot(., aes(x = bill_depth_mm, y = sex,fill = factor(stat(quantile)))) +
  stat_density_ridges(
    geom = "density_ridges_gradient", calc_ecdf = TRUE,
    quantiles = 4, quantile_lines = TRUE,
    jittered_points = TRUE,
    position = position_points_jitter(width = 0.05, height = 0),
    point_shape = '|', point_size = 3, point_alpha = 1) +
  scale_fill_viridis_d(name = "Quartiles")+
    theme_fivethirtyeight()+
  labs(title = 'Bill depth') + 
  facet_grid(species~.)


grid.arrange(p1,p2)

or another way would be to use facet_grid

penguins %>% 
  filter(!is.na(sex)) %>%
  pivot_longer(cols = bill_length_mm:bill_depth_mm, names_pattern =  'bill_(.*)_mm') %>% 
ggplot(., aes(x = value, y = sex,fill = factor(stat(quantile)))) +
  stat_density_ridges(
    geom = "density_ridges_gradient", calc_ecdf = TRUE,
    quantiles = 4, quantile_lines = TRUE,
    jittered_points = TRUE,
    position = position_points_jitter(width = 0.05, height = 0),
    point_shape = '|', point_size = 3, point_alpha = 1) +
  scale_fill_viridis_d(name = "Quartiles")+
    theme_hc()+
  labs(title = 'Bill size') + 
  facet_grid(species~name, scales = 'free')

Gapminder

Let’s see the Ratio girls/boys in school, primary and secondary education (%)

ratio_girls_education <- read_csv("../data/ratio_of_girls_to_boys_in_primary_and_secondary_education_perc.csv")

Parsed with column specification:
cols(
  .default = col_double(),
  country = col_character()
)
See spec(...) for full column specifications.

ratio_girls_education

let’s clean a little the data before we visualize it

ratio_girls_education <- ratio_girls_education %>% 
  pivot_longer(cols = -country,values_drop_na = TRUE, names_to = 'year', values_to = 'ratio') %>% 
    mutate(year = as.integer(year),
         country = as_factor(country),
         continent =  countrycode(sourcevar = country,
                            origin = "country.name",
                            destination = "continent")) %>% 
  left_join(gapminder_unfiltered, by = c("country", "year", "continent"))

We can plot one year

ratio_girls_education %>%
  filter(!is.na(lifeExp)) %>% 
  filter(year==2007) %>% 
  ggplot(aes(gdpPercap, ratio, color=continent,size=pop)) +
  geom_point()+
  annotate("rect", xmin=0, xmax=5000, ymin=-Inf, ymax=Inf, alpha=0.2, color="blue", fill="blue")+
  annotate("text", x = 13000, y = 0.7, label = "GDP per capita <5000")+
  theme_minimal()

plt <- ratio_girls_education %>%
  filter(!is.na(lifeExp)) %>% 
  filter(year==2007) %>% 
  ggplot(aes(gdpPercap, ratio, color=continent,size=pop, text = paste(country,
                                                                      '<br>GDP per capita: ',number(gdpPercap, 1),
                                                                      '<br>Population: ', number(pop/10000,1),'M',
                                                                      '<br>ratio: ',ratio))) +
  geom_point()+
  scale_y_continuous(labels= percent)+
  theme_minimal()


ggplotly(plt, tooltip = 'text')

Independent practice

Now design a plot yourselves!

1. Take a couple of minutes for selecting the dataset
2. Try to select an easy dataset (without too much pre-preprossesing), so you can try different visualizations of the data.

You can use any dataset you want, for example:

• other data from gapminder https://www.gapminder.org/data/

• tidytuesday is a project for dataviz that post a new dataset every week since 2018.

• Data from your own research (ideally)